search

sssssyf / SSMLP-RPL

Spectral-Spatial MLP Network with Reciprocal Points learning for Open-Set Hyperspectral Image Classification
11 stars 1 forks source link

I would like to have the code for this paper #1

Open wuyue197 opened 1 year ago

wuyue197 commented 1 year ago

I would like to ask for the code of your paper. Can you share it with me?

sssssyf commented 1 year ago

We are preparing the code and will release it as soon as possible, no more than two weeks.

wuyue197 commented 1 year ago

Thank you very much.

wuyue197 commented 1 year ago

I have a question about replicating code. The accuracy of the replicating open classes is only 2%. Why is that? Can you help me with that?

Neymar1011 commented 11 months ago

Hey buddy, I'm facing the same issue too. I'm only getting an accuracy of 1.7% when trying to reproduce the classification accuracy for unknown classes. Would you be open to discussing this and exchanging ideas?

I have a question about replicating code. The accuracy of the replicating open classes is only 2%. Why is that? Can you help me with that?

sssssyf commented 11 months ago

According to the previous experimental experience, increasing the patch size and trying different dynamic threshold schemes can increase the classification accuracy of unknown categories

Hey buddy, I'm facing the same issue too. I'm only getting an accuracy of 1.7% when trying to reproduce the classification accuracy for unknown classes. Would you be open to discussing this and exchanging ideas?

I have a question about replicating code. The accuracy of the replicating open classes is only 2%. Why is that? Can you help me with that?

ysqnb666 commented 10 months ago

s_sip_nods_s4': '/mnt/WXRC0020/users/junhao.zhang/tmp/slowfast/tools/s_sip_nods.pth Where can I download this file?